When life gives you confounders, uncover the hidden truths...
I have actively worked on developing the Python library pgmpy for over a year. My work involves implementing algorithms from the causal inference literature. As part of the European Summer of Code 2025, I worked on building tools for causal effect estimation, such as the DoubleML Regressor in a scikit-learn-compatible design. I have implemented several other algorithms, including the self-compatibility metric, Naive IV regressor, and IV identification methods. I have also previously mentored contributors in open-source programs, helping them work on algorithms such as causal identification and estimation methods.
Currently, I am building an influence diagram class to represent causal decision-making problems.
For my bachelor's thesis, I developed FermiGates, a novel algorithm for sparsifying neural networks that yields better predictions without any kind of regularisation.
I am now pursuing my Master's degree in Artificial Intelligence at University College London.
- Adds CID enhancement proposal - enhancement_proposals - Updated: 2026-10-16
- Self Compatibility score for the validation of causal discovery algorithms - pgmpy - Updated: 2026-10-12
- ports IV identification method - pgmpy - Updated: 2026-10-12
- [ENH] add _evalute and _evaluate_by_index methods for MSSE - sktime - Updated: 2025-08-25
- Add PINO layers and layer tests - deepchem - Updated: 2025-05-07
- Add Physics-Informed Neural Operator (PINO) model and unit tests - deepchem - Updated: 2025-05-03
- Add PINO model and unit tests - deepchem - Updated: 2025-05-02
- [DARHWOLF]: Add me as contributor - hacktoberfest-2022 - Updated: 2022-10-23
Last updated: 2026-10-18 03:04:41 UTC
- [ENH] ADD DoubleMLRegressor - pgmpy
- [ENH] ADD NaiveIVRegressor - pgmpy
- [ENH] Adds Base Causal Prediction Class - pgmpy
--